TECHNOLOGY FOR CHANGE

The Promises and Challenges of Exascale Computing

Salvatore Salamone | Date: 03-11-11 | Comments

Exascale computers will improve simulations, allow for more realistic modeling of complex natural processes, and offer better insight into the underlying science of those processes.

Table of Contents:

The Promises and Challenges of Exascale Computing
Funding for the development of exascale computing systems seems likely in the next federal budget. And while such systems will not be available for many years, researchers are already eyeing the benefits they will deliver.

In many fields�including life sciences, climate modeling and materials sciences�more powerful computers would help improve simulations, allow for more realistic modeling and offer better insight into the underlying science.

For example, many studies of physical phenomena use different scales depending on the scope of the simulation. A weather forecasting or climate model might use a fairly coarse grid when studying national or global patterns and turn to a very fine resolution model when studying localized phenomena such as lightning or cloud cover.

Similarly, a researcher studying the way a new chemical element interacts with a cell membrane might have one model that examines the interaction on a molecular level across a large area and another simulation that looks at atomic level interactions along a very small part of the molecule.

With more powerful computers, scientists could conduct their large-scale studies incorporating much more physics afforded by the more detailed simulations that examine smaller-scale phenomena.

The benefits of this approach was noted in a February ASCR Discovery Webzine article (ASCR is the U.S. Department of Energy's Office of Advanced Scientific Computing Research).

"One of the interesting things we want to do is global climate modeling�modeling the entire planet, but with resolution on a regional basis," said Mark Seager of the Lawrence Livermore National Laboratory. Such a regional model might be able to predict changes in the cycle that generates winter snows that melt to fill California's reservoirs.

"If the climate heats up to where precipitation comes down as rain instead of snow in the Sierras," Seager said, "then our water planning and infrastructure would need to radically change."

With current computing capabilities, scientists are not yet close to developing such models. The article noted that such models would require an increase in resolution of 10 to 100 times over current models. Adding such resolution to an entire global model is exactly the type of benefit an exascale computing system could deliver.

Next: Obstacles to Exascale >>